Pyrvinium Attenuates Hedgehog Signaling Downstream of Smoothened

Published OnlineFirst July 3, 2014; DOI: 10.1158/0008-5472.CAN-14-0317

Cancer Therapeutics, Targets, and Chemical Biology Research

Bin Li1, Dennis Liang Fei1, Colin A. Flaveny1, Nadia Dahmane2, Valerie Baubet2, Zhiqiang Wang1, Feng Bai1, Xin-Hai Pei1,3, Jezabel Rodriguez-Blanco1, Brian Hang4, Darren Orton5, Lu Han1, Baolin Wang6, Anthony J. Capobianco1,3,7, Ethan Lee4, and David J. Robbins1,3,7

Abstract The Hedgehog (HH) signaling pathway represents an important class of emerging developmental signaling pathways that play critical roles in the genesis of a large number of human cancers. The pharmaceutical industry is currently focused on developing small molecules targeting Smoothened (Smo), a key signaling effector of the HH pathway that regulates the levels and activity of the Gli family of transcription factors. Although one of these compounds, vismodegib, is now FDA-approved for patients with advanced basal cell carcinoma, acquired mutations in Smo can result in rapid relapse. Furthermore, many cancers also exhibit a Smo-independent activation of Gli proteins, an observation that may underlie the limited efﬁcacy of Smo inhibitors in clinical trials against other types of cancer. Thus, there remains a critical need for HH inhibitors with different mechanisms of action, particularly those that act downstream of Smo. Recently, we identiﬁed the FDA-approved anti-pinworm

compound pyrvinium as a novel, potent (IC50, 10 nmol/L) casein kinase-1a (CK1a) agonist. We show here that pyrvinium is a potent inhibitor of HH signaling, which acts by reducing the stability of the Gli family of transcription factors. Consistent with CK1a agonists acting on these most distal components of the HH signaling pathway, pyrvinium is able to inhibit the activity of a clinically relevant, vismodegib -resistant Smo mutant, as well as the Gli activity resulting from loss of the negative regulator suppressor of fused. We go on to demonstrate the utility of this small molecule in vivo, against the HH-dependent cancer medulloblastoma, attenuating its growth and reducing the expression of HH biomarkers. Cancer Res; 74(17); 4811–21. 2014 AACR.

Introduction Patched1 (Ptch1; ref. 2). Spontaneous cases of these tumors The Hedgehog (HH) signaling pathway plays key instruc- were subsequently shown to result from mutations or amplifi tional roles during embryonic development and adult tissue cations of a number of HH signaling components, including homeostasis. Consistent with this pivotal instructional role, Ptch1. Strong support for the role that HH signaling plays in the HH signaling pathway is commonly deregulated in many these cancers was provided from a number of genetic mouse human cancers (1). The role HH plays in cancer was first models of HH-driven medulloblastoma, in which mutations in identified in the inherited disorder Gorlin syndrome, which HH signaling components lead to the genesis of the same predisposes to basal cell carcinoma, medulloblastoma and tumors (3, 4). The growth of the tumors in these mice could also rhabdomyosarcoma, and results from loss-of-function muta- be abrogated by treatment with HH signaling inhibitors (5). tions in the gene encoding the HH core receptor component The pharmaceutical industry is currently focused on developing small molecules targeting Smoothened (Smo), a key signaling effector of the HH pathway that regulates the levels and activity of the Gli family of transcription factors (2). Although 1Molecular Oncology Program, Department of Surgery, University of Miami, Miami, Florida. 2Department of Neurosurgery, University of Penn- one of these compounds, vismodegib, is now FDA-approved for sylvania, Philadelphia, Pennsylvania. 3Sylvester Cancer Center, University patients with advanced basal cell carcinoma (2), acquired 4 of Miami, Miami, Florida. Department of Cell and Developmental Biology mutations in Smo can result in rapid relapse (6). Furthermore, and Vanderbilt Ingram Cancer Center, Vanderbilt University School of Medicine, Nashville, Tennessee. 5StemSynergy Therapeutics Inc., Miami, many cancers also exhibit a Smo-independent activation of Gli Florida. 6Weill Medical College, Cornell University, New York, New York. 7 proteins (7), an observation that may underlie the limited Department of Biochemistry and Molecular Biology, University of Miami, fi Miami, Florida. ef cacy observed for Smo inhibitors in clinical trials against other types of cancer (2). Thus, there remains a critical need for Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). HH inhibitors with different mechanisms of action, particularly those that act downstream of Smo. Corresponding Author: David J. Robbins, University of Miami Miller School of Medicine, The DeWitt Daughtry Family Department of Surgery, Molecular HH signaling is activated by binding of the HH ligands [Sonic Oncology Program, 1600 NW 10th Avenue. Miami, FL 33136. Phone: 305- (SHH), Indian, or Desert] to a receptor consisting of a Ptch 243-5717; Fax: 305-243-9694; E-mail: [email protected] protein (Ptch1 or Ptch2) and one of three coreceptors (8). This doi: 10.1158/0008-5472.CAN-14-0317 results in derepression of the G-protein–coupled seven-trans- 2014 American Association for Cancer Research. membrane protein Smo. Ultimately, canonical HH signaling

www.aacrjournals.org 4811

Li et al.

regulates the activity, proteolytic processing, and stability of Anthony Oro (Stanford University, Stanford, CA). Myc-Gli2 and members of the Gli family of transcription factors, Gli1-3 (7). GFP-Smo were from Addgene. Various Gli-null MEF cells were This regulation requires a number of protein kinases, including gifts of Dr. Wade Bushman (University of Wisconsin, Madison, protein kinase A (PKA), glycogen synthase kinase 3 (GSK3), and WI; ref. 26). shRNA constructs described here were purchased casein kinase1a (9–12), and the negative regulator suppressor from Open Biosystems, and used to prepare lentivirus as of fused (Sufu; refs. 13, 14). Mammalian HH signaling requires described therein. shRNA-expressing cell lines were selected trafficking through primary cilia, a membrane-encased micro- by 10 mg/mL puromycin. Smoothened agonist (SAG) and tubule-enriched organelle located on the apical side of polar- pyrvinium treatments were performed in the presence of ized cells (8, 15). Many components of the HH signaling 0.5% FBS. pathway transit through the primary cilia in their basal state and leave or enrich there in response to HH (8). During this Assays trafficking through the specialized environment of the primary Luciferase activity was determined using a luciferase detec- cilia, Gli proteins, likely through their interaction with Sufu, are tion kit (Promega), as previously described (27). Total RNA converted into their repressor forms or in response to HH from cells or tissues was extracted using the RNeasy kit converted into their active forms (16–18). In the basal state, (Qiagen), converted into cDNA (Applied Biosystems), then Gli2 and Gli3 are hyperphosphorylated at their Cul1-dependent analyzed using real-time RT-PCR and the cognate Taqman degrons (10, 11, 19). Subsequent to ubiquitination, Gli2 and probes, as per the manufacturer's instructions (Invitrogen). Gli3 are partially cleaved by proteasomes into their repressor Immunohistochemical analysis of Smo localization to primary forms. In response to HH, Gli2 and Gli3 become differentially cilia was performed as previously described (27). Immunohis- phosphorylated by PKA, at a distinct amino-terminal domain, tochemical staining of CK1a (Pierce) and Gli1 (27) were carried converting them into their activated nuclear forms (19–21). out using a Dako autostainer at the pathology core laboratory Nuclear-enriched Gli2 and Gli3 are labile and are quickly of the University of Miami. In vitro phosphorylation of recom- degraded by the proteasome through a Cul3-mediated ubiqui- binant human Gli1 by CK1a (Invitrogen) using radiolabeled tin proteasome system (17, 22). [32P]ATP was performed as previously described (28). Hexa- We recently reported that the FDA-approved drug pyrvi- histidine-tagged human Gli1 was produced using the Sf9/ nium is a novel and potent small molecule inhibitor of the Wnt baculovirus system. Statistical analysis was determined by the fi pathway (IC50, 10 nmol/L; ref. 23). We identi ed CK1a as the Student two-tailed t test, unless other stated. P values 0.05 critical cellular target of pyrvinium and showed that pyrvinium were considered statistically significant. acts as an allosteric activator of this protein kinase. Our in vitro and cellular binding studies demonstrated that pyrvinium Biochemistry binds avidly to CK1a (Kd, 1 nmol/L; ref. 23). Of the CK1 family Immunoblot analysis was carried out with the following members (a, g, d, and e), only CK1a is activated by pyrvinium. primary antibodies: anti-Gli1 (Cell Signaling Technology), anti- Pyrvinium has no effect on the activities of a panel of other Gli2 (R&D Systems), anti-Gli3 (9), anti-HA (Santa Cruz protein kinases representing all of the major branches of the Biotechnology), anti-Myc (Santa Cruz Biotechnology), anti- kinase superfamily (23), demonstrating that pyrvinium selec- Flag (Sigma), anti-Tubulin (Sigma), anti-Gapdh (Millipore), tively binds to and activates CK1a. As CK1a is implicated as a anti-ERK (Santa Cruz Biotechnology), anti-pAKT (Cell Signal- negative regulator of the HH signaling pathway, we hypothe- ing Technology), anti-pGSK3 (Cell Signaling Technology), anti- sized that pyrvinium might inhibit HH signaling (24). Here, we p-b-catenin (Cell Signaling Technology). CA-AKT plasmid show that pyrvinium does indeed attenuate HH signaling, and (myrAkt; Addgene), Bafilomycin A1 (Sigma), and MG132 (Cal- does so in vitro and in vivo—attenuating the growth of a well- biochem) were purchased. For immunoprecipitation, cells accepted HH-driven mouse tumor model. Furthermore, pyr- were lysed in a buffer containing 50 mmol/L Tris–HCl (pH, vinium acts to regulate Gli activity and stability downstream of 7.5), 150 mmol/L NaCl, 1 mmol/L EDTA, and 1% NP-40, Smo, in a CK1a-dependent manner, including attenuating HH supplemented with cOmplete Mini Protease Inhibitor Cocktail activity driven by a clinically relevant, vismodegib-resistant (Roche). After centrifugation, the supernatants were incubated Smo mutation (6). with the indicated antibodies overnight at 4C, and the immunoprecipitates were extensively washed with lysis buffer, elut- ed with SDS sample buffer, and boiled for 5 minutes before Materials and Methods analyses by immunoblotting. Cell culture NIH-3T3, HEK 293T, and Light-II cells were purchased from Mice and drug administration the American Type Culture Collection (ATCC) and were grown All mice were handled in accordance with the policies of the in medium as indicated by ATCC's instructions. Cerebellar University of Miami Institutional Animal Care and Use Com- þ granular precursor cells (GPC) were isolated as previously mittee. Spontaneous medulloblastomas from Ptch / mice described (25). NIH-3T3 cells stably expressing HA-Gli2 were (The Jackson Laboratory; Ptch1tm1Mps/J) were grafted onto a gift of Dr. Philip Beachy (Stanford University, Stanford, CA). CD-1 nude mice (Charles River Laboratories) subcutaneously. Transfections were performed using Lipofectamine 2000 (Invi- Drug treatment started when the tumors reached a size of trogen) according to the manufacturer's instructions. Myc- and approximately 100 mm3. For acute treatment, pyrvinium was Flag-tagged Gli1 and Myc-tagged b-Trcp were gifts of Dr. dissolved in DMSO and delivered through intraperitoneal

4812 Cancer Res; 74(17) September 1, 2014 Cancer Research

Casein Kinase 1a Agonists Attenuate HH Signaling

injection for the indicated times. For local peritumoral deliv- pyrvinium attenuated HH activity in a potent (IC50 of 10 ery, pyrvinium chloride was resuspended in PBS containing nmol/L), dose-dependent manner (Fig. 1A). Pyrvinium also 10% (2-Hydroxypropyl)-cyclodextrin and injected close to the inhibited HH-dependent expression of Gli1 and Ptch1, two tumor nodule every other day. well-established biomarkers of HH activity (Fig. 1B; ref. 1). In contrast, the inactive structural analog of pyrvinium VU211 (23) was unable to attenuate the expression of these HH Results biomarkers. We next evaluated the ability of pyrvinium to We recently identiﬁed CK1a as the critical cellular target of inhibit the proliferation of HH-dependent primary GPC, anoth- pyrvinium and showed that pyrvinium acts as a novel allosteric er well-established readout of HH activity (25, 29). Primary GPC activator of its protein kinase activity (23). As CK1a is also a key were isolated and treated with SHH plus pyrvinium or the Smo regulator of HH signaling (24), we hypothesized that pyrvinium antagonist cyclopamine. Pyrvinium attenuated GPC prolifer- would function as an inhibitor of this important developmen- ation, and did so in a manner similar to cyclopamine (Fig. 1C). tal signaling pathway. To test this hypothesis, we evaluated These data indicate that pyrvinium is a potent inhibitor of HH pyrvinium's ability to attenuate HH signaling. We stimulated signaling. HH activity in a HH reporter cell line (Light-II cells), using Pyrvinium has been reported to exhibit biologic effects via either SHH or the Smo agonist SAG, and in both cases mechanisms distinct from CK1a activation (30–33). Most

Figure 1. Pyrvinium suppresses HH signaling in a CK1a-dependent manner. A, Light-II cells treated with SHH, or the Smo agonist SAG, were subsequently incubated with the indicated doses of pyrvinium, or vehicle (DMSO), and luciferase activity was determined. B, SHH- treated NIH-3T3 cells were treated with DMSO, pyrvinium, or the inactive pyrvinium analog VU211. Twenty-four hours later, RNA was harvested from these cells and the expression of the HH target genes Gli1 and Ptch1 determined relative to that of the housekeeping gene Gapdh. C, primary GCP cells were incubated with BrdUrd and the indicated drugs (cyclopamine, 5 mmol/L; pyrvinium, 100 nmol/L) for 48 hours, followed by quantitation of BrdUrd incorporation using microscopy. D, Light-II cells were infected with lentivirus expressing the indicated shRNA. Seventy-two hours after infection, cells were treated with the indicated agents (SAG, 100 nmol/L; pyrvinium, 1, 10, 100 nmol/L) for 48 hours. RNA was harvested from these cells and the expression of Gli1 determined relative to the expression of Gapdh. E, Light-II cells were treated with the indicated agents in the presence or absence of the Wnt inhibitor IWR-1 (10 mmol/L) and luciferase activity was determined. F, HH or Wnt signaling activity in Gli2 / ; Gli3 / MEFs was induced with either SAG or Wnt3a conditioned media, in the absence or presence of 10 nmol/L pyrvinium. The expression of the Wnt target gene Dkk1, normalized to the expression of Gapdh, was analyzed as a readout of Wnt activity. Error bars, SEM (n ¼ 3); , P < 0.05.

www.aacrjournals.org Cancer Res; 74(17) September 1, 2014 4813

Li et al.

recently, it has been suggested that pyrvinium attenuates Wnt Pyrvinium potently inhibits HH signaling stimulated by the activity by blocking activation of AKT, preventing GSK3b oncogenic Smo-M2 mutant (data not shown). Given the emer- phosphorylation and its subsequent inactivation, which results gence of vismodegib resistance in the clinic (6), due to muta- in the phosphorylation and destabilization of b-catenin (34). tion of its binding site on Smo, we tested whether pyrvinium However, we show that the expression of a constitutive active could bypass a vismodegib-resistant Smo mutant (Smo-M2- AKT construct did not abolish pyrvinium's suppression of HH D473H; ref. 6) to inhibit HH signaling. Pyrvinium potently signaling (Supplementary Fig. S1A). Furthermore, we recently inhibited HH signaling activated by Smo-M2-D473H; vismo- determined pyrvinium's ability to bind to the active site of a degib had no observable effect (Fig. 2A). Thus, pyrvinium acts large panel of protein kinases (442 purified kinases) using to inhibit HH signaling via a mechanism distinct from vismo- Ambit's scanMAX kinase profiling service (data not shown). degib. We next determined the impact of pyrvinium treatment This screen examined the ability of pyrvinium to outcompete on the localization of Smo to the primary cilia of cells, which is kinases bound to a set of immobilized active site ligands. These an indication of Smo activation (35). Our results show that ligands bind to the active site of one or more kinases with high although pyrvinium treatment did result in a slight increase in fi af nity (Kd <1 nmol/L). Different concentrations of pyrvinium Smo localization, in the absence of SAG treatment, this were used to outcompete the individual kinases from the increase was not statistically significant (Fig. 2B). In addition, immobilized ligands. Our data suggest that pyrvinium does pyrvinium treatment had no impact on the increased Smo not bind to the active site of CK1a, as pyrvinium did not localization observed in response to activation by SAG, which compete with CK1a binding to its immobilized ligand in this suggests that pyrvinium acts downstream of Smo. Consistent assay. Consistent with this result, we did not observe the effects with this hypothesis, pyrvinium was able to inhibit constitutive of ATP on pyrvinium activity. We did, however, observe effects HH pathway activity in MEFs lacking Sufu, a negative regulator of pyrvinium on the conformation of CK1a as detected by an of HH signaling (14), whereas the Smo antagonist cyclopamine altered trypsin proteolysis pattern (23). Significantly, the scan- was unable to inhibit this activity (Fig. 2C). MAX kinase screen failed to detect inhibition of any of the Ultimately, HH signaling is mediated by Gli transcription kinases tested (using 1 mmol/L of pyrvinium), which included factors, whose activity and stability are regulated by a number AKT1-3 isoforms and all known PI3-kinases. Furthermore, of protein kinases, including CK1a (7). We, therefore, deter- consistent with pyrvinium acting through CK1a, we show that mined whether pyrvinium's ability to inhibit the expression of a the addition of pyrvinium to cells results in the time-dependent HH biomarker is Gli dependent. MEFs lacking Gli1, Gli2,orGli3 phosphorylation of a known CK1a substrate (Supplementary were all capable of responding to SAG in a manner that was Fig. S1B). attenuated by pyrvinium (Fig. 2D). Consistent with Gli2 and To more directly show that pyrvinium's ability to inhibit HH Gli3 playing redundant roles as HH-stimulated transcription signaling is CK1a dependent, we identified two shRNAs capa- factors, immortalized MEFs lacking both Gli2 and Gli3 did not ble of reducing CK1a protein levels (Supplementary Fig. S2A). respond to SAG to promote HH pathway activation (26). A low Light-II cells infected with this lentivirus were treated with level of SAG-induced activity was observed in MEFs lacking SAG to stimulate HH activity, followed by pyrvinium treat- both Gli1 and Gli2, and this low-level activation was mediated ment. CK1a knockdown reduced the capability of pyrvinium to by the activator form of Gli3 (26). Pyrvinium did not attenuate attenuate the expression of Gli1, relative to cells infected with this low level of activity, suggesting that pyrvinium does not control shRNA lentivirus (Fig. 1D). We also observed that regulate Gli3A-mediated transcriptional activity (Fig. 2D, knockdown of CK1a reduced the overall levels of SAG-induced right). Pyrvinium attenuated constitutive HH activity in cells HH target gene expression, as might be expected, given the overexpressing Gli1 (Fig. 2E) or Gli2 (data not shown), in positive role that CK1a also plays in Smo activation (35). contrast with the Smo antagonist vismodegib. These results The HH and Wnt signaling pathways negatively regulate or suggest that pyrvinium is capable of attenuating the transcrip- positively reinforce each other depending on cell context (36). tional activator function of Gli1 and Gli2, which are the major Thus, CK1a agonists could potentially attenuate HH signaling drivers of HH activity in cancer. indirectly via inhibition of Wnt signaling. To test this possi- As CK1a has been implicated in the proteolysis of the bility, we asked whether the Wnt inhibitors IWR-1 or XAV939, Drosophila Gli homolog Ci (8), we tested the hypothesis that which mediate their activity via mechanisms distinct from pyrvinium attenuates Gli activity by regulating its stability. We pyrvinium, similarly inhibit HH signaling. We found that IWR-1 transfected the NIH-3T3 cells with a plasmid expressing Gli1 and XAV939 did not suppress SAG-stimulated HH activation, and treated them with increasing doses of pyrvinium. Immu- nor did they potentiate the capacity of pyrvinium to inhibit noblotting of these cell lysates showed a dose-dependent HH signaling (Fig. 1E and data not shown). We then treated decrease in the levels of Gli1 relative to control (Fig. 3A). This Gli2 / ; Gli3 / immortalized mouse embryonic fibroblasts decrease in Gli1 protein levels occurred at doses of pyrvinium (MEF), which are not capable of responding to HH (26), with as low as 10 nmol/L, and was not a general consequence of Wnt3a in the presence and absence of pyrvinium. Although inhibiting HH activity because two structurally distinct Smo these MEFs did not respond to SAG, Wnt3a-induced signaling antagonists did not affect Gli1 protein levels (Fig. 3A, right). was still observed. Furthermore, Wnt signaling remained Similar results were obtained with Gli2, although in this case, pyrvinium sensitive in the absence of any HH signaling (Fig. pyrvinium-mediated destabilization was observed only when 1F). Thus, in these contexts, pyrvinium's ability to inhibit HH or HH signaling was activated (Fig. 3B). We also analyzed the Wnt signaling is independent of each other. levels of Gli3 protein in 3T3 cells treated with or without

4814 Cancer Res; 74(17) September 1, 2014 Cancer Research

Casein Kinase 1a Agonists Attenuate HH Signaling

Figure 2. Pyrvinium inhibits HH signaling downstream of Smo. A, cells expressing a vismodegib-resistant, oncogenic Smo mutant (D473-M2 Smo) were treated with the indicated drugs and then assayed for luciferase activity (left). The effectiveness of vismodegib was validated on Light-II cells treated with the Smo agonist SAG (right). B, NIH-3T3 cells stably expressing Smo–GFP were treated with vehicle, 100 nmol/L SAG, 100 nmol/L pyrvinium, or SAG plus pyrvinium, for 24 hours. Primary cilia localization of Smo was revealed by immunocytochemistry. Smo localization was manually quantitated over ﬁve random images. Representative images are shown at the bottom(green,Smo-GFP;blue,DAPI;red,primarycilia). þ C, Sufu / or Sufu / MEFs were treated with the indicated agents (5 mmol/Lcyclopamine)for3days.TheexpressionofGli1, Ptch1,and Hip was determined and normalized to that of Gapdh. D, the indicated MEFs were treated with the Smo agonist SAG in the presence or absence of pyrvinium (10 nmol/L). Twenty-four hours later, the expression of Gli1 (left) or Ptch1 (right) was determined. E, Light-II cells expressing Myc-Gli1 were treated with the indicated drugs, followed by quantitation of luciferase activity a day later. Error bars, SEM (n ¼ 3); , P < 0.05.

www.aacrjournals.org Cancer Res; 74(17) September 1, 2014 4815

Li et al.

Figure 3. Pyrvinium enhances the degradation of Gli transcription factors. A, NIH-3T3 cells expressing Flag-Gli1 were treated with the indicated agents (vismodegib, 100 nmol/L; cyclopamine, 5 mmol/L), followed by immunoblotting of these cellular lysates. B, NIH-3T3 cells expressing HA-Gli2 were treated with the indicated agents (SAG, 50 nmol/L; pyrvinium, 100 nmol/L), followed by immunoblotting of immunoprecipitated Gli2. C, NIH-3T3 were treated with indicated agents (SAG, 50 nmol/L; pyrvinium, 100 nmol/L), followed by immunoblotting of endogenous Gli3. Quantiﬁcation of multiple replicates (left) and a representative immunoblot (right) are shown for A–C. Error bars, SEM; NS, not statistically different; , P < 0.05. D, NIH-3T3 cells stably expressing the indicated shRNA were transfected with a plasmid expressing Myc-Gli1 (1 mg in control shRNA–infected cells and 0.5 mginCK1a-knockdown cells). These cells were subsequently treated with 10 or 100 nmol/L of pyrvinium, lysed, and then analyzed by immunoblotting. E, HEK 293T cells expressing Flag- Gli1 were treated with 200 nmol/L pyrvinium and 5 mmol/L of MG132 overnight. Anti-Flag immunoprecipitates, from lysates of these cells, were immunoblotted using the indicated antibodies. F, NIH-3T3 cells expressing Myc-Gli1 were treated with 100 mg/mL cycloheximide (CHX) at the indicated time points, in the presence or absence of 100 nmol/L pyrvinium. Immunoblotting was then performed to detect Gli1 and Gapdh levels. Error bars, SEM (n ¼ 3); , P < 0.05.

pyrvinium, and did not observe decreased levels of the full- know why activation of CK1a by pyrvinium is insufﬁcient to length Gli3 protein or increased levels of Gli3-R (Fig. 3C). In promote additional Gli3 processing, but speculate that the response to HH, both full-length Gli3 and Gli3-R levels were basal levels of CK1a activity are sufﬁcient to promote maxi- reduced regardless of pyrvinium treatment (Fig. 3C). This mum processing or that priming by PKA is the rate-limiting result is consistent with our analysis of pyrvinium's effect in event in Gli3 processing (19). Gli1 / ;Gli2 / double-knockout MEF cells, in which pyrvi- Consistent with pyrvinium acting through CK1a to regulate nium has no effect on HH signaling (Fig. 2D, right). We do not Gli stability, CK1a overexpression also resulted in destabilization

4816 Cancer Res; 74(17) September 1, 2014 Cancer Research

Casein Kinase 1a Agonists Attenuate HH Signaling

of the exogenous Gli1 protein (Supplementary Fig. S3A, left). In tion by pyrvinium is consistent with the specificity of this contrast, knockdown of CK1a attenuated pyrvinium's ability to association. Medulloblastoma tissue harboring a constitutively degrade Gli1 (Fig. 3D), whereas treatment with a CK1 antag- active HH signaling pathway coexpresses detectable levels of onist (Supplementary Fig. S3A, right) increased Gli protein endogenous Gli1 and CK1a (Fig. 4D and E). Using antibodies levels. To test the hypothesis that decreased stability of Glis specific to either of these proteins, we also observed an results from ubiquitin-mediated proteasomal and lysosomal- association between endogenous CK1a and endogenous Gli1 mediated degradation pathways, targeted via a CK1a-contain- from this tissue (Fig. 4E). The decreased stability of Glis that we ing protein complex, we applied pharmacologic inhibitors to observed in response to pyrvinium, combined with the asso- block specific protein degradation pathways and determined ciation between Gli proteins and CK1a, suggests that CK1a their effect on pyrvinium-mediated Gli destabilization. The may directly phosphorylate Glis to regulate their stability. proteasomal inhibitor MG-132, but not the lysosomal inhibitor Consistent with this suggestion, purified CK1a is capable of Bafilomycin A1, blocks pyrvinium-induced Gli1 destabilization phosphorylating purified Gli1, and this phosphorylation is (Supplementary Fig. S3B), suggesting that pyrvinium regulates further stimulated by pyrvinium (Supplementary Fig. S1C). Gli levels via proteasome-mediated degradation. Consistent To demonstrate the capacity of pyrvinium to attenuate the with this, pyrvinium greatly increased the level of Gli1 ubiqui- growth of HH activity–dependent cancers, we focused on a tination (Fig. 3E) and decreased the half-life of Gli1 protein in well-established mouse model of HH-dependent medulloblas- cells treated with the protein synthesis inhibitor cycloheximide toma that is routinely used to determine the efficacy of HH þ (Fig. 3F and Supplementary Fig. S3C). inhibitors, the Ptch / medulloblastoma model (2). We dis- þ We hypothesized that CK1a might regulate Gli stability sected a spontaneous medulloblastoma from a Ptch / mouse, through direct association, in a manner analogous to the and serially passaged the tumor as an allograft in CD-1 nude association described between the Drosophila Gli homolog Ci mice. We validated the dependency of HH signaling for these and CK1a (8). To test this possibility, we lysed HEK 293T cells tumors as HH (Gli1, Ptch2) but not Wnt (Axin2, Dkk1, and Lgr5) expressing Flag-tagged Gli1, Myc-tagged Gli2, or a control biomarkers were aberrantly activated (Fig. 5A). Acute treat- plasmid, immunoprecipitated either Gli1 or Gli2, and then ment of pyrvinium in this mouse model attenuated the expres- immunoblotted for associated CK1a. Endogenous CK1a asso- sion of the HH biomarkers Gli1 and Ptch2 (Fig. 5B). Chronic ciated with both Gli1 (Fig. 4A) and Gli2 (Fig. 4B), but was not treatment of such mice, using animals subcutaneously injected observed in the control immunoprecipitates. The association with pyrvinium adjacent to the approximately 100-mm3 between Gli1 and CK1a can be reduced following acute tumors, dramatically reduced the growth of medulloblastoma treatment with pyrvinium (Fig. 4C), and this decreased asso- allografts (Fig. 5C). Hematoxylin and eosin (H&E) staining ciation temporally preceded any observed Gli destabilization demonstrated that tumors from pyrvinium-treated animals (data not shown). The reversibility of the Gli1/CK1a associa- showed decreased cancer cells and increased fibrotic tissue,

Figure 4. CK1a associates with Gli transcription factors. A and B, HEK A B Input IP Input IP 293T cells transfected with Ctrl - - + + + - - pcDNA3.1-Flag, Flag-Gli1 (A) or Flag-Gli1 - + - + Ctrl + Myc-Gli2 - + - + pcDNA3, Myc-Gli2 (B) were IP: Flag subjected to immunoprecipitation 37 kDa IP: Myc 37 kDa IB:CK1 a a and immunoblotting with the IB: CK1 IP: Flag IP: Myc indicated antibodies. C, 150 kDa 250 kDa IB: Flag IB: Myc pcDNA3.1-Flag (Ctrl) or Flag-Gli1– transfected HEK 293T cells were treated with vehicle or 200 nmol/L Medulloblastoma pyrvinium for 1 hour, before any IP C D a detectable changes in Gli1 protein Ctrl + - - Gli1 CK1 levels. Samples were Flag-Gli1 - + + immunoprecipitated with Flag M2 Pyrvinium --+ beads and subject to immunoblot IP: Flag 37 kDa IHC analysis for CK1a or Flag-Gli1. IB: CK1a IP: Flag D, immunohistologic staining of 150 kDa CK1a and Gli1 proteins in IB: Flag þ Ptch / -derived medulloblastoma tissue. Scale bar, 100 mm. E, endogenous CK1a (left) or Gli1 E (right) was immunoprecipitated from homogenized input IP: IgGIP: Gli1 input Ip: IgGIp: CK1a medulloblastoma tissue using the Gli1 150 kDa Gli1 150 kDa indicated antisera (2 mg). These IgG HC IgG HC immunoprecipitates were then CK1a 37 kDa CK1a 37 kDa immunoblotted for CK1a or Gli1. IgG LC

www.aacrjournals.org Cancer Res; 74(17) September 1, 2014 4817

Li et al.

ABMedulloblastoma 20 * * Gli1 1.4 1.4 Ptch2 15 1.2 1.2 Axin2 1.0 1.0 Dkk1 0.8 0.8 10 Lgr5 0.6 0.6 0.4 0.4 expression expression Relative Gli1

5 Relative Ptch2 expression 0.2 0.2 Relative gene 0.0 0.0 h 0 le 2 ic 8 h icle 12 h h m 8 h 1 N1 N2 T1 T2 h m u m e u m Ve i V ni iu in i n v v r viniu yr rvi y r P y P P Py

C D Medulloblastoma Medulloblastoma Vehicle Vehicle Pyrvinium 1,000 Pyrvinium ) 3 800 # 600 * H&E 400

200 * Tumor volume (mm 0 1357911 *, Fewer cancer cells; #, Fibrosis Days

E NS NS * * 2.5 1.4 1.4 2.5

1.2 1.2 2.0 2.0 1.0 1.0 1.5 1.5 0.8 0.8 0.6 0.6 1.0 1.0 expression expression 0.4 0.4 expression expression Relative Lgr5

Relative Gli1 0.5 Relative Ptch2 0.5 Relative Axin2 0.2 0.2 0.0 0.0 0.0 0.0 le m le ic ium hic n eh i Vehicle V Vehicle Ve rv rviniu y Pyrvinium P Pyrvinium Py

þ Figure 5. Pyrvinium attenuates the growth of a Ptch / -derived medulloblastoma allograft. A, RNA was extracted from normal mouse brains (N1 or N2) or þ Ptch / -derived medulloblastoma tissue (T1 or T2). Real-time RT-PCR was then used to detect the expression of the indicated HH or Wnt target þ genes, normalized to 18S ribosomal RNA. B, allografts of Ptch / -derived medulloblastoma were grown in nude mice until they were approximately 100 mm3 in size. These mice were subsequently treated with 5 mg/kg pyrvinium, or vehicle, for the indicated times. The expression of Gli1 or Ptch2 was then quantitated by real-time RT-PCR. Error bars, SEM (n ¼ 5); , P < 0.05. C, allografts of similar medulloblastoma were grown in nude mice until they were approximately 100 mm3 in size. These mice were subsequently treated with vehicle or pyrvinium (0.8 mg/kg) by subcutaneous injection every 2 days. Tumor volumes were measured at the indicated time points. Error bars, SEM (n ¼ 6); , P < 0.05. D, representative H&E staining of tumors from vehicle and pyrvinium-treated þ mice in C. Scale bar, 100 mm. E, following the chronic administration of pyrvinium in mice harboring Ptch / medulloblastoma allografts in C, tumors were harvested and subjected to real time RT-PCR analysis for HH or Wnt target genes. Error bars, SEM (n ¼ 6); , P < 0.05.

compared with that from the vehicle-treated controls (Fig. 5D), tumor cells, in which a reduction in total gene expression consistent with a reduction in the size of the tumor driven by might be expected. Thus, the CK1a agonist pyrvinium attenu- decreased numbers of tumor cells. In the pyrvinium-treated ates HH pathway activity and tumor growth in vivo in a well- tumors, the expression of HH target genes was decreased established model of HH-driven medulloblastoma. relative to the vehicle control group (Fig. 5E). However, Wnt target genes, which are not hyperactivated in this subtype of medulloblastoma (see Fig. 5A), were not different between Discussion pyrvinium and vehicle-treated tissue (Fig. 5E). These observa- We show here that the FDA-approved anthelmintic drug tions are inconsistent with a generalized, nonspeciﬁc effect on pyrvinium can be repurposed as a potent HH inhibitor.

4818 Cancer Res; 74(17) September 1, 2014 Cancer Research

Casein Kinase 1a Agonists Attenuate HH Signaling

Consistent with pyrvinium acting as a CK1a agonist (23), we recognition that drugs that act via allosteric mechanisms are show that it attenuates HH signaling in a CK1a-dependent not constrained by active site chemistry and are much more manner. We further show that CK1a and Gli proteins associate selective (sites may not be conserved even between proteins in and propose a model by which pyrvinium induces Gli desta- the same family) in their mechanism of action. bilization downstream of Smo. Unlike vismodegib, the only CK1a participates in a number of cellular processes other Smo inhibitor currently FDA approved (2), pyrvinium is able to than HH and Wnt signaling, including retinoid X receptor attenuate increased Gli activity resulting from loss of Sufu, regulation (43). Interestingly, similar to its role in HH signaling, overexpression of a Gli protein, or a Smo protein harboring a CK1a plays a dual function in NF-kB signaling, both promoting clinically relevant vismodegib-resistant mutation. Pyrvinium and repressing receptor-induced NF-kB activity (44). However, was also able to attenuate the growth of a well-established HH- this is unlikely to be an impediment for its further development þ driven cancer model, a Ptch / -driven medulloblastoma allo- as a target for treating cancer. In fact, inhibitors that target graft mouse model (5, 37). Interestingly, a significant number of other distinct CK1 isoforms (d and e) are currently in preclin- human medulloblastomas result from loss of Sufu or amplifi- ical development in both academia and industry, to treat a cation of Gli2 (1), and thus would be resistant to vismodegib variety of human pathologies (45, 46). In addition, there is but might conceivably respond to a CK1a agonist such as significant precedence for the development of very effective pyrvinium. In its current method of dosing and formulation, anticancer agents that target proteins involved in basic cellular pyrvinium likely lacks the pharmacokinetic properties to be processes, which takes advantage of a tumor's increased used in the clinic against HH-driven tumors. However, sensitivity to such agents. Many of these have already been improvements in its formulation or delivery, such as perhaps approved by the FDA or are in clinical trials. These include direct ventricular delivery into the brain, might rapidly be paclitaxel (microtubule stabilizer), vorinostat (HDAC inhibi- adapted to treat such patients with late-stage cancer for whom tor), velcade (proteasome inhibitor), temsirolimus (mTOR few therapeutic options remain. inhibitor), and geldanamycin derivatives (targets HSP90 and Importantly, given the emergence of noncanonical Gli acti- currently in phase II clinical trials). vation in human cancers (1), pyrvinium's effects on HH signaling occurred downstream of Smo, where it induces the Disclosure of Potential Conflicts of Interest destabilization of Gli proteins. A number of other Gli inhibitors D. Orton received a commercial research grant from and has ownership interest (including patents) in StemSynergy Therapeutics, Inc. E. Lee is the have now been described (2). For example, although Gli founder of and has ownership interest (including patents) in StemSynergy antagonists (GANT) are now widely used Gli inhibitors in the Therapeutics, Inc. D.J. Robbins has ownership interest (including patents) in and is a consultant/advisory board member for StemSynergy Therapeutics, Inc. laboratory setting, their IC50s are more than 100-times higher No potential conflicts of interest were disclosed by the other authors. than the CK1a agonists described here. Similar potency con- cerns also exist for another class of HH inhibitor HPI (IC50,10 mmol/L), which perturb Gli processing and stability. Arsenic Authors' Contributions Conception and design: B. Li, D.L. Fei, C.A. Flaveny, D. Orton, A.J. Capobianco, trioxide is another FDA-approved drug that has been repur- D.J. Robbins posed as a Gli inhibitor in vitro and in vivo. Although this drug is Development of methodology: B. Li, D.L. Fei, C.A. Flaveny, F. Bai, A.J. Capobianco already clinically available, its limited potency and well- Acquisition of data (provided animals, acquired and managed patients, described dose-limiting toxicities in humans, which occur at provided facilities, etc.): B. Li, D.L. Fei, C.A. Flaveny, N. Dahmane, Z. Wang, doses similar to those required to inhibit HH signaling, may X.-H. Pei, B. Hang, L. Han, A.J. Capobianco, E. Lee Analysis and interpretation of data (e.g., statistical analysis, biostatistics, limit its usefulness in the clinic. computational analysis): B. Li, D.L. Fei, C.A. Flaveny, N. Dahmane, J. Rodri- Although CK1 was initially thought to be constitutively guez-Blanco, B. Hang, D. Orton, L. Han, A.J. Capobianco, E. Lee, D.J. Robbins active, recent evidence suggests that individual members of Writing, review, and/or revision of the manuscript: B. Li, C.A. Flaveny, N. Dahmane, D. Orton, A.J. Capobianco, E. Lee, D.J. Robbins this protein kinase family are regulated via an array of tran- Administrative, technical, or material support (i.e., reporting or orga- scriptional and posttranslational mechanisms (38). For CK1a, nizing data, constructing databases): B. Li, C.A. Flaveny, V. Baubet, D. Orton, A.J. Capobianco, D.J. Robbins such mechanisms include the frequent silencing of CK1a in Study supervision: B. Li, C.A. Flaveny, A.J. Capobianco, D.J. Robbins melanoma and loss-of-heterozygosity of the CK1a gene in ap- Other (provided Gli3 antibody): B. Wang proximately 30% of all human tumors (http://www.sanger.ac. uk/cgi-bin/genetics/CGP/cghviewer/CghHome.cgi)—consis- Acknowledgments tent with CK1a activity being growth suppressing in many The authors thank Drs. Karoline Briegel and Teresa Zimmers and members of the Robbins, Capobianco, and Lee laboratories for providing insight during fi human cancers (39, 40). The identi cation of pyrvinium as an discussions about this article. The authors also thank Dr. Rune Toftgard allosteric activator of CK1a led to the speculation that pyrvi- (Karolinska Institute) for providing Sufu / MEF cells. nium may be mimicking an endogenous regulator of CK1a. This was confirmed with the recent identification of a novel Grant Support family of proteins that regulate the activity of CK1 kinases This work was supported by the following grants: NIH: CA082628, GM64011, a GM103926, GM081635, and P50CA95103; Alex Lemonade Stand Foundation, and in vivo, and the demonstration that CK1 activity can be The University of Miami Women's Cancer Association. attenuated by extracellular ligands, such as Wnts (41, 42). The costs of publication of this article were defrayed in part by the payment of These newly identified levels of regulation could potentially page charges. This article must therefore be hereby marked advertisement in be usurped for the development of novel anticancer therapeu- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. tics that target individual CK1 isoforms. The excitement of a Received February 5, 2014; revised May 19, 2014; accepted June 4, 2014; class of anticancer agents such as pyrvinium comes from the published OnlineFirst July 3, 2014.

www.aacrjournals.org Cancer Res; 74(17) September 1, 2014 4819

Li et al.

References 1. Teglund S, Toftgard R. Hedgehog beyond medulloblastoma and basal 23. Thorne CA, Hanson AJ, Schneider J, Tahinci E, Orton D, Cselenyi CS, cell carcinoma. Biochim Biophys Acta 2010;1805:181–208. et al. Small-molecule inhibition of Wnt signaling through activation of 2. Amakye D, Jagani Z, Dorsch M. Unraveling the therapeutic potential of casein kinase 1alpha. Nat Chem Biol 2010;6:829–36. the Hedgehog pathway in cancer. Nat Med 2013;19:1410–22. 24. Lum L, Yao S, Mozer B, Rovescalli A, Von Kessler D, Nirenberg M, et al. 3. Goodrich LV, Milenkovic L, Higgins KM, Scott MP. Altered neural cell Identification of Hedgehog pathway components by RNAi in Drosoph- fates and medulloblastoma in mouse patched mutants. Science ila cultured cells. Science 2003;299:2039–45. 1997;277:1109–13. 25. Dahmane N, Sanchez P, Gitton Y, Palma V, Sun T, Beyna M, et al. The 4. Zibat A, Uhmann A, Nitzki F, Wijgerde M, Frommhold A, Heller T, et al. Sonic Hedgehog-Gli pathway regulates dorsal brain growth and Time-point and dosage of gene inactivation determine the tumor tumorigenesis. Development 2001;128:5201–12. spectrum in conditional Ptch knockouts. Carcinogenesis 2009;30: 26. Lipinski RJ, Bijlsma MF, Gipp JJ, Podhaizer DJ, Bushman W. Estab- 918–26. lishment and characterization of immortalized Gli-null mouse embry- 5. Berman DM, Karhadkar SS, Hallahan AR, Pritchard JI, Eberhart CG, onic fibroblast cell lines. BMC Cell Biol 2008;9:49. Watkins DN, et al. Medulloblastoma growth inhibition by hedgehog 27. Fei DL, Li H, Kozul CD, Black KE, Singh S, Gosse JA, et al. Activation of pathway blockade. Science 2002;297:1559–61. Hedgehog signaling by the environmental toxicant arsenic may con- 6. Yauch RL, Dijkgraaf GJ, Alicke B, Januario T, Ahn CP, Holcomb T, et al. tribute to the etiology of arsenic-induced tumors. Cancer Res 2010;70: Smoothened mutation confers resistance to a Hedgehog pathway 1981–8. inhibitor in medulloblastoma. Science 2009;326:572–4. 28. Cselenyi CS, Jernigan KK, Tahinci E, Thorne CA, Lee LA, Lee E. LRP6 7. Hui CC, Angers S. Gli proteins in development and disease. Annu Rev transduces a canonical Wnt signal independently of Axin degradation Cell Dev Biol 2011;27:513–37. by inhibiting GSK3's phosphorylation of beta-catenin. Proc Natl Acad 8. Robbins DJ, Fei DL, Riobo NA. The Hedgehog signal transduction Sci U S A 2008;105:8032–7. network. Sci Signal 2012;5:re6. 29. Wechsler-Reya RJ, Scott MP. Control of neuronal precursor pro- 9. Wang B, Fallon JF, Beachy PA. Hedgehog-regulated processing of liferation in the cerebellum by Sonic Hedgehog. Neuron 1999;22: Gli3 produces an anterior/posterior repressor gradient in the devel- 103–14. oping vertebrate limb. Cell 2000;100:423–34. 30. Shen M, Bellaousov S, Hiller M, de La Grange P, Creamer TP, Malina O, 10. Pan Y, Bai CB, Joyner AL, Wang B. Sonic hedgehog signaling reg- et al. Pyrvinium pamoate changes alternative splicing of the serotonin ulates Gli2 transcriptional activity by suppressing its processing and receptor 2C by influencing its RNA structure. Nucleic Acids Res 2013; degradation. Mol Cell Biol 2006;26:3365–77. 41:3819–32. 11. Tempe D, Casas M, Karaz S, Blanchet-Tournier MF, Concordet JP. 31. Tomitsuka E, Kita K, Esumi H. An anticancer agent, pyrvinium pamoate Multisite protein kinase A and glycogen synthase kinase 3beta phos- inhibits the NADH-fumarate reductase system–a unique mitochondrial phorylation leads to Gli3 ubiquitination by SCFbetaTrCP. Mol Cell Biol energy metabolism in tumour microenvironments. J Biochem 2012; 2006;26:4316–26. 152:171–83. 12. Bhatia N, Thiyagarajan S, Elcheva I, Saleem M, Dlugosz A, Mukhtar H, 32. Harada Y, Ishii I, Hatake K, Kasahara T. Pyrvinium pamoate inhibits et al. Gli2 is targeted for ubiquitination and degradation by beta-TrCP proliferation of myeloma/erythroleukemia cells by suppressing mito- ubiquitin ligase. J Biol Chem 2006;281:19320–6. chondrial respiratory complex I and STAT3. Cancer Lett 2012;319: 13. Chen MH, Wilson CW, Li YJ, Law KK, Lu CS, Gacayan R, et al. 83–8. Cilium-independent regulation of Gli protein function by Sufu in 33. Yu DH, Macdonald J, Liu G, Lee AS, Ly M, Davis T, et al. Pyrvinium Hedgehog signaling is evolutionarily conserved. Genes Dev 2009; targets the unfolded protein response to hypoglycemia and its anti- 23:1910–28. tumor activity is enhanced by combination therapy. PLoS ONE 2008;3: 14. Svard J, Heby-Henricson K, Persson-Lek M, Rozell B, Lauth M, e3951. Bergstrom A, et al. Genetic elimination of Suppressor of fused reveals 34. Venerando A, Girardi C, Ruzzene M, Pinna LA. Pyrvinium pamoate an essential repressor function in the mammalian Hedgehog signaling does not activate protein kinase CK1, but promotes Akt/PKB down- pathway. Dev Cell 2006;10:187–97. regulation and GSK3 activation. Biochem J 2013;452:131–7. 15. Rohatgi R, Milenkovic L, Scott MP. Patched1 regulates hedgehog 35. Chen Y, Sasai N, Ma G, Yue T, Jia J, Briscoe J, et al. Sonic Hedgehog signaling at the primary cilium. Science 2007;317:372–6. dependent phosphorylation by CK1alpha and GRK2 is required for 16. Humke EW, Dorn KV, Milenkovic L, Scott MP, Rohatgi R. The output ciliary accumulation and activation of smoothened. PLoS Biol 2011;9: of Hedgehog signaling is controlled by the dynamic association e1001083. between Suppressor of Fused and the Gli proteins. Genes Dev 36. Hooper JE, Scott MP. Communicating with Hedgehogs. Nat Rev Mol 2010;24:670–82. Cell Biol 2005;6:306–17. 17. Wen X, Lai CK, Evangelista M, Hongo JA, de Sauvage FJ, Scales SJ. 37. Wong H, Alicke B, West KA, Pacheco P, La H, Januario T, et al. Kinetics of hedgehog-dependent full-length Gli3 accumulation in Pharmacokinetic-pharmacodynamic analysis of vismodegib in pre- primary cilia and subsequent degradation. Mol Cell Biol 2010;30: clinical models of mutational and ligand-dependent Hedgehog path- 1910–22. way activation. Clin Cancer Res 2011;17:4682–92. 18. Tukachinsky H, Lopez LV, Salic A. A mechanism for vertebrate Hedge- 38. Cheong JK, Virshup DM. Casein kinase 1: complexity in the family. Int J hog signaling: recruitment to cilia and dissociation of SuFu-Gli protein Biochem Cell Biol 2011;43:465–9. complexes. J Cell Biol 2010;191:415–28. 39. Elyada E, Pribluda A, Goldstein RE, Morgenstern Y, Brachya G, 19. Wang B, Li Y. Evidence for the direct involvement of {beta}TrCP Cojocaru G, et al. CKIalpha ablation highlights a critical role for p53 in Gli3 protein processing. Proc Natl Acad Sci U S A 2006;103: in invasiveness control. Nature 2011;470:409–13. 33–8. 40. Sinnberg T, Menzel M, Kaesler S, Biedermann T, Sauer B, Nahnsen S, 20. PanY,WangC,WangB.PhosphorylationofGli2byproteinkinase et al. Suppression of casein kinase 1alpha in melanoma cells induces a A is required for Gli2 processing and degradation and the Sonic switch in beta-catenin signaling to promote metastasis. Cancer Res Hedgehog-regulated mouse development. Dev Biol 2009;326: 2010;70:6999–7009. 177–89. 41. Cruciat CM, Dolde C, de Groot RE, Ohkawara B, Reinhard C, 21. Niewiadomski P, Kong JH, Ahrends R, Ma Y, Humke EW, Khan S, et al. Korswagen HC, et al. RNA helicase DDX3 is a regulatory subunit Gli protein activity is controlled by multisite phosphorylation in verte- of casein kinase 1 in Wnt-beta-catenin signaling. Science 2013; brate hedgehog signaling. Cell Rep 2014;6:168–81. 339:1436–41. 22. Zhang Q, Shi Q, Chen Y, Yue T, Li S, Wang B, et al. Multiple Ser/Thr-rich 42. Hernandez AR, Klein AM, Kirschner MW. Kinetic responses of degrons mediate the degradation of Ci/Gli by the Cul3-HIB/SPOP E3 beta-catenin specify the sites of Wnt control. Science 2012;338: ubiquitin ligase. Proc Natl Acad Sci U S A 2009;106:21191–6. 1337–40.

4820 Cancer Res; 74(17) September 1, 2014 Cancer Research

Casein Kinase 1a Agonists Attenuate HH Signaling

43. ZhaoY,QinS,AtanganLI,MolinaY,OkawaY,ArpawongHT,etal. 45. Arey R, McClung CA. An inhibitor of casein kinase 1 epsilon/delta Casein kinase 1alpha interacts with retinoid X receptor and inter- partially normalizes the manic-like behaviors of the ClockDelta19 feres with agonist-induced apoptosis. J Biol Chem 2004;279: mouse. Behav Pharmacol 2012;23:392–6. 30844–9. 46. Behrend L, Milne DM, Stoter M, Deppert W, Campbell LE, Meek DW, 44. Bidere N, Ngo VN, Lee J, Collins C, Zheng L, Wan F, et al. Casein kinase et al. IC261, a speciﬁc inhibitor of the protein kinases casein kinase 1alpha governs antigen-receptor-induced NF-kappaB activation and 1-delta and -epsilon, triggers the mitotic checkpoint and induces human lymphoma cell survival. Nature 2009;458:92–6. p53-dependent postmitotic effects. Oncogene 2000;19:5303–13.

www.aacrjournals.org Cancer Res; 74(17) September 1, 2014 4821

Pyrvinium Attenuates Hedgehog Signaling Downstream of Smoothened

Bin Li, Dennis Liang Fei, Colin A. Flaveny, et al.

Cancer Res 2014;74:4811-4821. Published OnlineFirst July 3, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-14-0317

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2014/07/21/0008-5472.CAN-14-0317.DC1

Cited articles This article cites 46 articles, 24 of which you can access for free at: http://cancerres.aacrjournals.org/content/74/17/4811.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/74/17/4811.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/74/17/4811. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.